The first frame proved the core could compose layers. The next question was what those layers could become.
Two things mattered to me here. Effects had to belong to the layer that needs them, not to a global setting somewhere. And the old problem that started all this, placing an image convincingly at an angle, had to be solved properly.
Effects that belong to layers
In the tools I had used, effects were scarce and attached to inputs. If two scenes used the same camera differently, you worked around it.
In Muxivo Studio an effect is part of a layer. The same source can appear in two scenes, keyed in one and colour-corrected in the other, and each layer keeps its own chain. Effects run on the GPU as part of the composition, in the order you set.
The first set covered what a live show needs every day:
- Chroma key and luma key, for green screens and graphics with a black background.
- Colour correction, so sources from different cameras sit together.
- Shadows with perspective, so a cut-out presenter or object casts a shadow that follows the angle of the scene instead of a flat drop shadow.
A few days later the layer manager could stack several effects on a layer and reorder them. That small detail changes how you work: you build a look step by step and see each step live.
The seam, revisited
Remember the thin line across a distorted image, the bug that never got fixed? The underlying need was simple: put a picture into a scene at a convincing angle, as if a screen were standing in the set.
Doing that by hand is guesswork. You drag four corners until it looks about right, and it never quite matches the room, because the room was filmed through a real lens from a real position.
So instead of guessing the corners, Muxivo Studio recovers the camera.
A camera from a photograph
Here is the idea. Take a photo or a frame of the set. Mark a few lines that are parallel in the real world, the edges of a table, a wall, the floor. From those lines it is possible to work out where the camera stood, where it pointed and how wide its lens was.
Muxivo Studio imports that camera from fSpy, a free tool built exactly for this. The result is a 3D scene with a perspective camera that matches the photo: its position, its target, its field of view.
Then comes the part that makes it useful for live production. A normal 2D scene can be bound to that 3D scene. Its layers are projected through the matched camera, so a video placed on a wall sits on that wall, in the right perspective, without dragging corners by eye.
The 3D workspace itself is a practical editor: a viewport, simple objects such as planes, boxes, spheres and screens, and transforms to place them. These act as guides for staging and projection.
What this is, and what it is not yet
This is not a virtual production system. There is no full 3D rendering of the set, no tracked camera, no materials or lighting. Those belong to later phases, and I would rather say so plainly.
What it is, is the first plank of the bridge I wrote about in the first chapter. You stay in the 2D workflow you know: scenes, layers, effects. Space arrives gradually, starting with a camera that agrees with the room.
What comes next
Layers, effects and a sense of space make a good picture. They do not make a show. A show needs things that come and go on top of the picture: titles, lower thirds, transitions, stingers.
How those should work turned out to be one of the most important decisions in the whole project. That is the next chapter.